Harrison's Neurology in Clinical Medicine, 3rd Edition


Charles A. Czeisler Image John W. Winkelman Image Gary S. Richardson

Disturbed sleep is among the most frequent health complaints physicians encounter. More than one-half of adults in the United States experience at least intermittent sleep disturbance. For most, it is an occasional night of poor sleep or daytime sleepiness. However, the Institute of Medicine has estimated that 50–70 million Americans suffer from a chronic disorder of sleep and wakefulness, which can lead to serious impairment of daytime functioning. In addition, such problems may contribute to or exacerbate medical or psychiatric conditions. Thirty years ago, many such complaints were treated with hypnotic medications without further diagnostic evaluation. Since then, a distinct class of sleep and arousal disorders has been identified.


Given the opportunity, most adults will sleep 7–8 h per night, although the timing, duration, and internal structure of sleep vary among healthy individuals and as a function of age. At the extremes, infants and the elderly have frequent interruptions of sleep. In the United States, adults tend to have one consolidated sleep episode per day, although in some cultures sleep may be divided into a mid-afternoon nap and a shortened night sleep. Two principal neural systems govern the expression of the sleep and wakefulness states within the daily cycle. The first potentiates sleep in proportion to the duration of wakefulness (the “sleep homeostat”), while the second rhythmically modulates sleep and wakefulness tendencies at appropriate phases of the 24-h day (the circadian clock). Intrinsic abnormalities in the function of either of these systems, or extrinsic disturbances (environmental, drug- or illness-related) that supersede their normal expression, can lead to clinically recognizable sleep disorders.


States and stages of human sleep are defined on the basis of characteristic patterns in the electroencephalogram (EEG), the electrooculogram (EOG—a measure of eye-movement activity), and the surface electro-myogram (EMG) measured on the chin and neck. The continuous recording of this array of electrophysiologic parameters to define sleep and wakefulness is termed polysomnography.

Polysomnographic profiles define two states of sleep: (1) rapid-eye-movement (REM) sleep and (2) non-rapid-eye-movement (NREM) sleep. NREM sleep is further subdivided into three stages, characterized by increasing arousal threshold and slowing of the cortical EEG. REM sleep is characterized by a low-amplitude, mixed-frequency EEG similar to that of NREM stage N1 sleep. The EOG shows bursts of REM similar to those seen during eyes-open wakefulness. Chin EMG activity is absent, reflecting the brainstem-mediated muscle atonia that is characteristic of that state.


Normal nocturnal sleep in adults displays a consistent organization from night to night (Fig. 20-1). After sleep onset, sleep usually progresses through NREM stages N1–N3 sleep within 45–60 min. Slow-wave sleep (NREM stage N3 sleep) predominates in the first third of the night and comprises 15–25% of total nocturnal sleep time in young adults. The percentage of slow-wave sleep is influenced by several factors, most notably age (see later). Prior sleep deprivation increases the rapidity of sleep onset and both the intensity and amount of slow-wave sleep.

The first REM sleep episode usually occurs in the second hour of sleep. More rapid onset of REM sleep in an adult (particularly if <30 min) may suggest pathology such as endogenous depression, narcolepsy, circadian rhythm disorders, or drug withdrawal. NREM and REM alternate through the night with an average period of 90–110 min (the “ultradian” sleep cycle). Overall, REM sleep constitutes 20–25% of total sleep, and NREM stages N1 and N2 are 50–60%.



Stages of REM sleep (solid bars), the three stages of NREM sleep, and wakefulness over the course of the entire night for representative young and older adult men. Characteristic features of sleep in older people include reduction of slow-wave sleep, frequent spontaneous awakenings, early sleep onset, and early morning awakening. (From the Division of Sleep Medicine, Brigham and Women’s Hospital.)

Age has a profound impact on sleep state organization (Fig. 20-1). Slow-wave sleep is most intense and prominent during childhood, decreasing sharply coincident with puberty and across the second and third decades of life. After age 30, there is a continued decline in the amount of slow-wave sleep, and the amplitude of delta EEG activity comprising slow-wave sleep is profoundly reduced. The depth of slow-wave sleep, as measured by the arousal threshold to auditory stimulation, also decreases with age. In the otherwise healthy older person, slow-wave sleep may be completely absent, particularly in males. Paradoxically, older people are better able to tolerate acute sleep deprivation than young adults, maintaining reaction time and sustaining vigilance with fewer lapses of attention.

A different age profile exists for REM sleep than for slow-wave sleep. In infancy, REM sleep may comprise 50% of total sleep time, and the percentage is inversely proportional to developmental age. The amount of REM sleep falls off sharply over the first postnatal year as a mature REM-NREM cycle develops; thereafter, REM sleep occupies a relatively constant percentage of total sleep time.


Experimental studies in animals have variously implicated the medullary reticular formation, the thalamus, and the basal forebrain in the generation of sleep, while the brainstem reticular formation, the midbrain, the subthalamus, the thalamus, and the basal forebrain have all been suggested to play a role in the generation of wakefulness or EEG arousal.

Current models suggest that the capacity for sleep and wakefulness generation is distributed along an axial “core” of neurons extending from the brainstem rostrally to the basal forebrain. A cluster of γ-aminobutyric acid (GABA) and galaninergic neurons in the ventrolateral preoptic (VLPO) hypothalamus is selectively activated coincident with sleep onset. These neurons project to and inhibit the multiple neural wakefulness centers that comprise the ascending arousal system, and selective cell-specific lesions of VLPO substantially reduce sleep time, indicating that the hypothalamic VLPO neurons play an executive role in sleep regulation. More recent data have identified another sleep center, the median preoptic nucleus (MnPOn) of the hypothalamus with similar activation patterns and projections, suggesting that, like that of wakefulness, executive control of sleep may also be multicentric.

Specific regions in the pons are associated with the neurophysiologic correlates of REM sleep. Small lesions in the dorsal pons result in the loss of the descending muscle inhibition normally associated with REM sleep; microinjections of the cholinergic agonist carbachol into the pontine reticular formation produce a state with all of the features of REM sleep. These experimental manipulations are mimicked by pathologic conditions in humans and animals. A prominent feature of narcolepsy, for example, is abrupt, complete, or partial paralysis (cataplexy) in response to a variety of stimuli, a pathologic activation of neural systems mediating the atonia of normal REM sleep. In narcoleptic dogs, physostigmine, a central cholinesterase inhibitor, increases the frequency of cataplexy episodes, while atropine decreases their frequency. Conversely, in REM sleep behavior disorder (see later), patients suffer from a failure of normal motor inhibition during REM sleep, resulting in involuntary, occasionally violent, movement arising out of dream episodes.


Early experimental studies that focused on the raphe nuclei of the brainstem appeared to implicate serotonin as the primary sleep-promoting neurotransmitter, while catecholamines were considered to be responsible for wakefulness. Simple neurochemical models have given way to more complex formulations involving multiple parallel waking systems. Pharmacologic studies suggest that histamine, acetylcholine, dopamine, serotonin, and noradrenaline are all involved in wake promotion. A novel neuropeptide, orexin (also known as hypocretin), localized to a cluster of neurons in the lateral hypothalamus and originally identified based on its pathogenic role in narcolepsy (see later), also appears to be involved in the control of wakefulness.

In the basal forebrain (BF), adenosine receptors on cholinergic neurons are thought to play a role in assessing homeostatic sleep need by providing an index of cellular energy status. Projections from these BF neurons to executive sleep centers such as VLPO thus allow incorporation of homeostatic sleep need into the control of sleep state expression. At a practical level, this model suggests that the alerting effects of caffeine, an adenosine receptor antagonist, reflect the attenuation of the homeostatic signal from the basal forebrain.

The prominent hypnotic effects of benzodiazepine receptor agonists suggest that endogenous ligands of this receptor may be involved in normal sleep physiology. While neurosteroids with activity at this receptor have been identified, their role in normal sleep-wake control remains unclear. In addition, a broad array of endogenous sleep- and wake-promoting substances have been identified, the role of which in normal sleep-wake control remains unclear. These include corticotropin-releasing hormone (CRH), prostaglandin D2, delta sleep–inducing peptide, muramyl dipeptide, interleukin 1, fatty acid primary amides, and melatonin.


The sleep-wake cycle is the most evident of the many 24-h rhythms in humans. Prominent daily variations also occur in endocrine, thermoregulatory, cardiac, pulmonary, renal, gastrointestinal, and neurobehavioral functions. At the molecular level, endogenous circadian rhythmicity is driven by self-sustaining transcriptional/translational feedback loops (Fig. 20-2). In evaluating a daily variation in humans, it is important to distinguish between those rhythmic components passively evoked by periodic environmental or behavioral changes (e.g., the increase in blood pressure and heart rate that occurs upon assumption of the upright posture) and those actively driven by an endogenous oscillatory process (e.g., the circadian variation in plasma cortisol that persists under a variety of environmental and behavioral conditions).



Model of the molecular feedback loop at the core of the mammalian circadian clock. The positive element of the feedback loop (+) is the transcriptional activation of the Per1 gene (and probably other clock genes) by a heterodimer of the transcription factors CLOCK and BMAL1 (also called MOP3) bound to an E-box DNA regulatory element. The Per1 transcript and its product, the clock component PER1 protein, accumulate in the cell cytoplasm. As it accumulates, the PER1 protein is recruited into a multiprotein complex thought to contain other circadian clock component proteins such as cryptochromes (CRYs), Period proteins (PERs), and others. This complex is then transported into the cell nucleus (across the dotted line), where it functions as the negative element in the feedback loop (–) by inhibiting the activity of the CLOCK-BMAL1 transcription factor heterodimer. As a consequence of this action, the concentration of PER1 and other clock proteins in the inhibitory complex falls, allowing CLOCK-BMAL1 to activate transcription of Per1 and other genes and begin another cycle. The dynamics of the 24-h molecular cycle are controlled at several levels, including regulation of the rate of PER protein degradation by casein kinase-1 epsilon (CK1E). Additional limbs of this genetic regulatory network, omitted for the sake of clarity, are thought to contribute stability. Question marks denote putative clock proteins, such as Timeless (TIM), as yet lacking genetic proof of a role in the mammalian clock mechanism. (Copyright Charles J. Weitz, PhD, Department of Neurobiology, Harvard Medical School.)

While it is now recognized that many peripheral tissues in mammals have circadian clocks that regulate diverse physiologic processes, these independent tissue-specific oscillations are coordinated by a central neural pacemaker located in the suprachiasmatic nuclei (SCN) of the hypothalamus. Bilateral destruction of these nuclei results in a loss of the endogenous circadian rhythm of locomotor activity, which can be restored only by transplantation of the same structure from a donor animal. The genetically determined period of this endogenous neural oscillator, which averages ~24.2 h in humans, is normally synchronized to the 24-h period of the environmental light-dark cycle. Small differences in circadian period underlie variations in diurnal preference, with the circadian period shorter in individuals who typically rise early compared to those who typically go to bed late. Entrainment of mammalian circadian rhythms by the light-dark cycle is mediated via the retinohypothalamic tract, a monosynaptic pathway that links specialized, photoreceptive retinal ganglion cells directly to the SCN. Humans are exquisitely sensitive to the resetting effects of light, particularly the shorter wavelengths (~460–500 nm) of the visible spectrum.

The timing and internal architecture of sleep are directly coupled to the output of the endogenous circadian pacemaker. Paradoxically, the endogenous circadian rhythms of sleep tendency, sleepiness, and REM sleep propensity all peak near the habitual wake time, just after the nadir of the endogenous circadian temperature cycle, whereas the circadian wake propensity rhythm peaks 1–3 h before the habitual bedtime. These rhythms are thus timed to oppose the homeostatic decline of sleep tendency during the habitual sleep episode and the rise of sleep tendency throughout the usual waking day, respectively. Misalignment of the output of the endogenous circadian pacemaker with the desired sleep-wake cycle can, therefore, induce insomnia, decreased alertness, and impaired performance evident in night-shift workers and airline travelers.


Polysomnographic staging of sleep correlates with behavioral changes during specific states and stages. During the transitional state between wakefulness and sleep (stage N1 sleep), subjects may respond to faint auditory or visual signals without “awakening.” Short-term memory incorporation is inhibited at the onset of NREM stage N1 sleep, which may explain why individuals aroused from that transitional sleep stage frequently deny having been asleep. Such transitions may intrude upon behavioral wakefulness after sleep deprivation, notwithstanding attempts to remain continuously awake (see “Shift-Work Disorder,” later in this chapter).

Awakenings from REM sleep are associated with recall of vivid dream imagery >80% of the time. The reliability of dream recall increases with REM sleep episodes occurring later in the night. Imagery may also be reported after NREM sleep interruptions, though these typically lack the detail and vividness of REM sleep dreams. The incidence of NREM sleep dream recall can be increased by selective REM sleep deprivation, suggesting that REM sleep and dreaming per se are not inexorably linked.


All major physiologic systems are influenced by sleep. Changes in cardiovascular function include a decrease in blood pressure and heart rate during NREM and particularly during slow-wave sleep. During REM sleep, phasic activity (bursts of eye movements) is associated with variability in both blood pressure and heart rate mediated principally by the vagus nerve. Cardiac dysrhythmias may occur selectively during REM sleep. Respiratory function also changes. In comparison to relaxed wakefulness, respiratory rate becomes more regular during NREM sleep (especially slow-wave sleep) and tonic REM sleep and becomes very irregular during phasic REM sleep. Minute ventilation decreases in NREM sleep out of proportion to the decrease in metabolic rate at sleep onset, resulting in a higher PCO2.

Endocrine function also varies with sleep. Slow-wave sleep is associated with secretion of growth hormone in men, while sleep in general is associated with augmented secretion of prolactin in both men and women. Sleep has a complex effect on the secretion of luteinizing hormone (LH): during puberty, sleep is associated with increased LH secretion, whereas sleep in the post-pubertal female inhibits LH secretion in the early follicular phase of the menstrual cycle. Sleep onset (and probably slow-wave sleep) is associated with inhibition of thyroid-stimulating hormone and of the adrenocorticotropic hormone–cortisol axis, an effect that is superimposed on the prominent circadian rhythms in the two systems.

The pineal hormone melatonin is secreted predominantly at night in both day- and night-active species, reflecting the direct modulation of pineal activity by the circadian pacemaker through a circuitous neural pathway from the SCN to the pineal gland. Melatonin secretion is not dependent upon the occurrence of sleep, persisting in individuals kept awake at night. Secretion is inhibited by ambient light, an effect mediated by a neural connection from the retina via the SCN. The role of endogenous melatonin in normal sleep-wake regulation is unclear, but administration of exogenous melatonin can potentiate sleepiness and facilitate sleep onset when administered in the afternoon or evening, at a time when endogenous melatonin levels are low. The efficacy of melatonin as a sleep-promoting therapy for patients with insomnia is currently not known.

Sleep is also accompanied by alterations of thermoregulatory function. NREM sleep is associated with an attenuation of thermoregulatory responses to either heat or cold stress, and animal studies of thermosensitive neurons in the hypothalamus document an NREM-sleep-dependent reduction of the thermoregulatory set-point. REM sleep is associated with complete absence of thermoregulatory responsiveness, resulting in functional poikilothermy. However, the potential adverse impact of this failure of thermoregulation is blunted by inhibition of REM sleep by extreme ambient temperatures.



PATIENT Sleep Disorders

Patients may seek help from a physician because of one of several symptoms: (1) an acute or chronic inability to initiate or maintain sleep adequately at night (insomnia); (2) chronic fatigue, sleepiness, or tiredness during the day; or (3) a behavioral manifestation associated with sleep itself. The specific approach to an insomnia complaint will depend on the nature of comorbid medical or psychiatric disease, if present. In general, however, the insomnia complaint should be specifically addressed as soon as it is recognized. This more aggressive approach reflects growing evidence that chronic insomnia may contribute to comorbid disease processes. For example, specific management of symptomatic insomnia at the time of diagnosis of major depressive disorder (MDD) has been shown to positively impact the response to antidepressants. Evidence that insomnia and sleep loss affect the perception of pain suggests that a similar approach is warranted in acute and chronic pain management. In general, at least for chronic insomnia, there is little evidence justifying an expectant approach in which specific insomnia therapy is deferred while comorbid disease is first addressed.

Table 20-1 outlines the diagnostic and therapeutic approach to the patient with a complaint of excessive daytime sleepiness.

TABLE 20-1



A careful history is essential. In particular, the duration, severity, and consistency of the symptoms are important, along with the patient’s estimate of the consequences of the sleep disorder on waking function. Information from a friend or family member can be invaluable; some patients may be unaware of, or will underreport, such potentially embarrassing symptoms as heavy snoring or falling asleep while driving.

Patients with excessive sleepiness should be advised to avoid all driving until effective therapy has been achieved.

Completion by the patient of a day-by-day sleep-work-drug log for at least 2 weeks can help the physician better understand the nature of the complaint. Work times and sleep times (including daytime naps and nocturnal awakenings) as well as drug and alcohol use, including caffeine and hypnotics, should be noted each day.

Polysomnography is necessary for the diagnosis of specific disorders such as narcolepsy and sleep apnea and may be of utility in other settings as well.


Insomnia is the complaint of inadequate sleep; it can be classified according to the nature of sleep disruption and the duration of the complaint. Insomnia is subdivided into difficulty falling asleep (sleep onset insomnia), frequent or sustained awakenings (sleep maintenance insomnia), or early morning awakenings (sleep offset insomnia), though most insomnia patients present with two or more of these symptoms. Other insomnia patients present with persistent sleepiness/fatigue despite sleep of adequate duration (nonrestorative sleep). Similarly, the duration of the symptom influences diagnostic and therapeutic considerations. An insomnia complaint lasting one to several nights (within a single episode) is termed transient insomnia and is typically the result of situational stress or a change in sleep schedule or environment (e.g., jet lag disorder). Short-term insomnia lasts from a few days to 3 weeks. Disruption of this duration is usually associated with more protracted stress, such as recovery from surgery or short-term illness. Long-term insomnia, or chronic insomnia, lasts for months or years and, in contrast with short-term insomnia, requires a thorough evaluation of underlying causes (see later in this chapter). Chronic insomnia is often a waxing and waning disorder, with spontaneous or stressor-induced exacerbations.

An occasional night of poor sleep, typically in the setting of stress or excitement about external events, is both common and without lasting consequences. However, persistent insomnia can lead to impaired daytime function, injury due to accidents, and the development of major depression. In addition, there is emerging evidence that individuals with chronic insomnia have increased utilization of health care resources, even after controlling for comorbid medical and psychiatric disorders.

All insomnias can be exacerbated and perpetuated by behaviors that are not conducive to initiating or maintaining sleep. Inadequate sleep hygiene is characterized by a behavior pattern prior to sleep or a bedroom environment that is not conducive to sleep, or irregularity in the timing or duration of the nightly sleep episode. Noise, light, or technology (e.g., television, radio, cell phone, mobile email device or computer) in the bedroom can interfere with sleep, as can a bed partner with periodic limb movements during sleep or one who snores loudly. Clocks can heighten the anxiety about the time it has taken to fall asleep. Drugs that act on the central nervous system, large meals, vigorous exercise, or hot showers just before sleep may all interfere with sleep onset. Many individuals participate in stressful work-related activities in the evening, producing a state incompatible with sleep onset. In preference to hypnotic medications, patients should be counseled to avoid stressful activities before bed, develop a soporific bedtime ritual, and prepare and reserve the bedroom environment for sleeping. Consistent, regular bedtimes and rising times should be maintained daily, including weekends.


Many patients with chronic insomnia have no clear, single identifiable underlying cause for their difficulties with sleep. Rather, such patients often have multiple etiologies for their insomnia, which may evolve over the years. In addition, the chief sleep complaint may change over time, with initial insomnia predominating at one point, and multiple awakenings or nonrestorative sleep occurring at other times. Subsyndromal psychiatric disorders (e.g., anxiety and mood complaints), negative conditioning to the sleep environment (psychophysiologic insomnia, see next), amplification of the time spent awake (paradoxical insomnia), physiologic hyperarousal, and poor sleep hygiene (see earlier) may all be present. As these processes may be both causes and consequences of chronic insomnia, many individuals will have a progressive course to their symptoms in which the severity is proportional to the chronicity, and much of the complaint may persist even after effective treatment of the initial inciting etiology. Treatment of insomnia is often directed to each of the putative contributing factors: behavior therapies for anxiety and negative conditioning (see later), pharmacotherapy and/or psychotherapy for mood/anxiety disorders, and an emphasis on maintenance of good sleep hygiene.

If insomnia persists after treatment of these contributing factors, empirical pharmacotherapy is often used on a nightly or intermittent basis. A variety of sedative compounds are used for this purpose. Alcohol and anti-histamines are the most commonly used nonprescription sleep aids. The former may help with sleep onset but is associated with sleep disruption during the night and can escalate into abuse, dependence, and withdrawal in the predisposed individual. Antihistamines, which are the primary active ingredient in most over-the-counter sleep aids, may be of benefit when used intermittently but often produce rapid tolerance and may have multiple side effects (especially anticholinergic), which limit their use, particularly in the elderly. Benzodiazepine-receptor agonists are the most effective and well-tolerated class of medications for insomnia. The broad range of half-lives allows flexibility in the duration of sedative action. The most commonly prescribed agents in this family are zaleplon (5–20 mg), with a half-life of 1–2 h; zolpidem (5–10 mg) and triazolam (0.125–0.25 mg), with half-lives of 2–3 h; eszopiclone (1–3 mg), with a half-life of 5.5–8 h; and temazepam (15–30 mg) and lorazepam (0.5–2 mg), with half-lives of 6–12 h. Generally, side effects are minimal when the dose is kept low and the serum concentration is minimized during the waking hours (by using the shortest-acting effective agent). At least one benzodiazepine receptor agonist (eszopiclone) continues to be effective for 6 months of nightly use. However, longer durations of use have not been evaluated, and it is unclear whether this is true of other agents in this class. Moreover, with even brief continuous use of benzodiazepine receptor agonists, rebound insomnia can occur upon discontinuation. The likelihood of rebound insomnia and tolerance can be minimized by short durations of treatment, intermittent use, or gradual tapering of the dose. For acute insomnia, nightly use of a benzodiazepine receptor agonist for a maximum of 2–4 weeks is advisable. For chronic insomnia, intermittent use is recommended, unless the consequences of untreated insomnia outweigh concerns regarding chronic use. Benzodiazepine receptor agonists should be avoided, or used very judiciously, in patients with a history of substance or alcohol abuse. The heterocyclic antidepressants (trazodone, amitriptyline, and doxepin) are the most commonly prescribed alternatives to benzodiazepine receptor agonists due to their lack of abuse potential and lower cost. Trazodone (25–100 mg) is used more commonly than the tricyclic antidepressants as it has a much shorter half-life (5–9 h), has much less anticholinergic activity (sparing patients, particularly the elderly, constipation, urinary retention, and tachycardia), is associated with less weight gain, and is much safer in overdose. The risk of priapism is small (~1 in 10,000).

Psychophysiologic insomnia

Persistent psychophysiologic insomnia is a behavioral disorder in which patients are preoccupied with a perceived inability to sleep adequately at night. This sleep disorder begins like any other acute insomnia; however, the poor sleep habits and sleep-related anxiety (“insomnia phobia”) persist long after the initial incident. Such patients become hyperaroused by their own efforts to sleep or by the sleep environment, and the insomnia becomes a conditioned or learned response. Patients may be able to fall asleep more easily at unscheduled times (when not trying) or outside the home environment. Polysomnographic recording in patients with psychophysiologic insomnia reveals an objective sleep disturbance, often with an abnormally long sleep latency; frequent nocturnal awakenings; and an increased amount of stage N1 transitional sleep. Rigorous attention should be paid to improving sleep hygiene, correction of counterproductive, arousing behaviors before bedtime, and minimizing exaggerated beliefs regarding the negative consequences of insomnia. Behavioral therapies are the treatment modality of choice, with intermittent use of medications. When patients are awake for >20 min, they should read or perform other relaxing activities to distract themselves from insomnia-related anxiety. In addition, bedtime and wake time should be scheduled to restrict time in bed to be equal to their perceived total sleep time. This will generally produce sleep deprivation, greater sleep drive, and, eventually, better sleep. Time in bed can then be gradually expanded. In addition, methods directed toward producing relaxation in the sleep setting (e.g., meditation, muscle relaxation) are encouraged.

Adjustment insomnia (acute insomnia)

This typically develops after a change in the sleeping environment (e.g., in an unfamiliar hotel or hospital bed) or before or after a significant life event, such as a change of occupation, loss of a loved one, illness, or anxiety over a deadline or examination. Increased sleep latency, frequent awakenings from sleep, and early morning awakening can all occur. Recovery is generally rapid, usually within a few weeks. Treatment is symptomatic, with intermittent use of hypnotics and resolution of the underlying stress. Altitude insomnia describes a sleep disturbance that is a common consequence of exposure to high altitude. Periodic breathing of the Cheyne-Stokes type occurs during NREM sleep about half the time at high altitude, with restoration of a regular breathing pattern during REM sleep. Both hypoxia and hypocapnia are thought to be involved in the development of periodic breathing. Frequent awakenings and poor quality sleep characterize altitude insomnia, which is generally worse on the first few nights at high altitude but may persist. Treatment with acetazol-amide can decrease time spent in periodic breathing and substantially reduce hypoxia during sleep.


Insomnia associated with mental disorders

Approximately 80% of patients with psychiatric disorders describe sleep complaints. There is considerable heterogeneity, however, in the nature of the sleep disturbance both between conditions and among patients with the same condition. Depression can be associated with sleep onset insomnia, sleep maintenance insomnia, or early morning wakefulness. However, hypersomnia occurs in some depressed patients, especially adolescents and those with either bipolar or seasonal (fall/winter) depression (Chap. 54). Indeed, sleep disturbance is an important vegetative sign of depression and may commence before any mood changes are perceived by the patient. Consistent polysomnographic findings in depression include decreased REM sleep latency, lengthened first REM sleep episode, and shortened first NREM sleep episode; however, these findings are not specific for depression, and the extent of these changes varies with age and symptomatology. Depressed patients also show decreased slow-wave sleep and reduced sleep continuity.

In mania and hypomania, sleep latency is increased and total sleep time can be reduced. Patients with anxiety disorders tend not to show the changes in REM sleep and slow-wave sleep seen in endogenously depressed patients. Chronic alcoholics lack slow-wave sleep, have decreased amounts of REM sleep (as an acute response to alcohol), and have frequent arousals throughout the night. This is associated with impaired daytime alertness. The sleep of chronic alcoholics may remain disturbed for years after discontinuance of alcohol usage. Sleep architecture and physiology are disturbed in schizophrenia, with a decreased amount of slow-wave sleep (NREM stage N3 sleep) and a lack of augmentation of REM sleep following REM sleep deprivation; chronic schizophrenics often show day-night reversal, sleep fragmentation, and insomnia.

Insomnia associated with neurologic disorders

A variety of neurologic diseases result in sleep disruption through both indirect, nonspecific mechanisms (e.g., pain in cervical spondylosis or low back pain) or by impairment of central neural structures involved in the generation and control of sleep itself. For example, dementia from any cause has long been associated with disturbances in the timing of the sleep-wake cycle, often characterized by nocturnal wandering and an exacerbation of symptomatology at night (so-called sundowning).

Epilepsy may rarely present as a sleep complaint (Chap. 26). Often the history is of abnormal behavior, at times with convulsive movements during sleep. The differential diagnosis includes REM sleep behavior disorder, sleep apnea syndrome, and periodic movements of sleep (see earlier in this chapter). Diagnosis requires nocturnal polysomnography with a full EEG montage. Other neurologic diseases associated with abnormal movements, such as Parkinson’s diseasehemiballismusHuntington’s chorea, and Tourette’s syndrome (Chap. 30), are also associated with disrupted sleep, presumably through secondary mechanisms. However, the abnormal movements themselves are greatly reduced during sleep. Headache syndromes (migraine or cluster headache) may show sleep-associated exacerbations (Chap. 8) by unknown mechanisms.

Fatal familial insomnia is a rare hereditary disorder caused by degeneration of anterior and dorsomedial nuclei of the thalamus. Insomnia is a prominent early symptom. Patients develop progressive autonomic dysfunction, followed by dysarthria, myoclonus, coma, and death. The pathogenesis is a mutation in the prion gene (Chap. 43).

Insomnia associated with other medical disorders

A number of medical conditions are associated with disruptions of sleep. The association is frequently nonspecific, e.g., sleep disruption due to chronic pain from rheumatologic disorders. Attention to this association is important in that sleep-associated symptoms are often the presenting or most bothersome complaint. Treatment of the underlying medical problem is the most useful approach. Sleep disruption can also result from the use of medications such as glucocorticoids (see later).

One prominent association is between sleep disruption and asthma. In many asthmatics there is a prominent daily variation in airway resistance that results in marked increases in asthmatic symptoms at night, especially during sleep. In addition, treatment of asthma with theophylline-based compounds, adrenergic agonists, or glucocorticoids can independently disrupt sleep. When sleep disruption is a side effect of asthma treatment, inhaled glucocorticoids (e.g., beclomethasone) that do not disrupt sleep may provide a useful alternative.

Cardiac ischemia may also be associated with sleep disruption. The ischemia itself may result from increases in sympathetic tone as a result of sleep apnea. Patients may present with complaints of nightmares or vivid, disturbing dreams, with or without awareness of the more classic symptoms of angina or of the sleep-disordered breathing. Treatment of the sleep apnea may substantially improve the angina and the nocturnal sleep quality. Paroxysmal nocturnal dyspnea can also occur as a consequence of sleep-associated cardiac ischemia that causes pulmonary congestion exacerbated by the recumbent posture.

Chronic obstructive pulmonary disease is also associated with sleep disruption, as are cystic fibrosismenopausehyperthyroidismgastroesophageal refluxchronic renal failure, and liver failure.


Disturbed sleep can result from ingestion of a wide variety of agents. Caffeine is perhaps the most common pharmacologic cause of insomnia. It produces increased latency to sleep onset, more frequent arousals during sleep, and a reduction in total sleep time for up to 8–14 h after ingestion. Even small amounts of coffee can significantly disturb sleep in some patients; therefore, a 1- to 2-month trial without caffeine should be attempted in patients with these symptoms. Similarly, alcohol and nicotine can interfere with sleep, despite the fact that many patients use them to relax and promote sleep. Although alcohol can increase drowsiness and shorten sleep latency, even moderate amounts of alcohol increase awakenings in the second half of the night. In addition, alcohol ingestion prior to sleep is contraindicated in patients with sleep apnea because of the inhibitory effects of alcohol on upper airway muscle tone. Acutely, amphetamines and cocaine suppress both REM sleep and total sleep time, which return to normal with chronic use. Withdrawal leads to an REM sleep rebound. A number of prescribed medications can produce insomnia. Antidepressants, sympathomimetics, and glucocorticoids are common causes. In addition, severe rebound insomnia can result from the acute withdrawal of hypnotics, especially following the use of high doses of benzodiazepines with a short half-life. For this reason, hypnotic doses should be low to moderate, and prolonged drug tapering is encouraged.


Patients with this sensorimotor disorder report an irresistible urge to move the legs, or sometimes the upper extremities, that is often associated with creepy-crawling or aching dysesthesias deep within the affected limbs. For most patients with RLS, the dysesthesias and restlessness are much worse in the evening or night compared to the daytime and frequently interfere with the ability to fall asleep. The symptoms appear with inactivity and are temporarily relieved by movement. In contrast, paresthesias secondary to peripheral neuropathy persist with activity. The severity of this chronic disorder may wax and wane over time and can be exacerbated by sleep deprivation, caffeine, alcohol, serotonergic antidepressants, and pregnancy. The prevalence is 1–5% of young to middle-age adults and 10–20% of those aged >60 years. There appear to be important differences in RLS prevalence among racial groups, with higher prevalence in those of Northern European ancestry. Roughly one-third of patients (particularly those with an early age of onset) will have multiple affected family members. At least three separate chromosomal loci have been identified in familial RLS, though no gene has been identified to date. Iron deficiency and renal failure may cause RLS, which is then considered secondary RLS. The symptoms of RLS are exquisitely sensitive to dopaminergic drugs (e.g., pramipexole 0.25–0.5 mg q8PM or ropinirole 0.5–4 mg q8PM), which are the treatments of choice. Opioids, benzodiazepines, and gabapentin may also be of therapeutic value. Most patients with restless legs also experience periodic limb movements of sleep, although the reverse is not the case.


Periodic limb movements of sleep (PLMS), previously known as nocturnal myoclonus, consists of stereotyped, 0.5- to 5.0-s extensions of the great toe and dorsiflexion of the foot, which recur every 20–40 s during NREM sleep, in episodes lasting from minutes to hours, as documented by bilateral surface EMG recordings of the anterior tibialis on polysomnography. PLMS is the principal objective polysomnographic finding in 17% of patients with insomnia and 11% of those with excessive daytime somnolence (Fig. 20-3). It is often unclear whether it is an incidental finding or the cause of disturbed sleep. When deemed to be the latter, PLMS is called PLMD. PLMS occurs in a wide variety of sleep disorders (including narcolepsy, sleep apnea, REM sleep behavior disorder, and various forms of insomnia) and may be associated with frequent arousals and an increased number of sleep-stage transitions. The pathophysiology is not well understood, though individuals with high spinal transections can exhibit periodic leg movements during sleep, suggesting the existence of a spinal generator. Treatment options include dopaminergic medications or benzodiazepines.



Polysomnographic recordings of (A) obstructive sleep apnea and (B) periodic limb movement of sleep. Note the snoring and reduction in air flow in the presence of continued respiratory effort, associated with the subsequent oxygen desaturation (upper panel). Periodic limb movements occur with a relatively constant intermovement interval and are associated with changes in the EEG and heart rate acceleration (lower panel). RAT, right anterior tibialis; LAT, left anterior tibialis. (From the Division of Sleep Medicine, Brigham and Women’s Hospital.)


Daytime impairment due to sleep loss may be difficult to quantify for several reasons. First, patients may be unaware of the extent of sleep deprivation. In obstructive sleep apnea, for example, the brief arousals from sleep associated with respiratory recovery after each apneic episode results in daytime sleepiness, despite the fact that the patient may be unaware of the sleep fragmentation. Second, subjective descriptions of waking impairment vary from patient to patient. Patients may describe themselves as “sleepy,” “fatigued,” or “tired” and may have a clear sense of the meaning of those terms, while others may use the same terms to describe a completely different condition. Third, sleepiness, particularly when profound, may affect judgment in a manner analogous to ethanol, such that subjective awareness of the condition and the consequent cognitive and motor impairment is reduced. Finally, patients may be reluctant to admit that sleepiness is a problem, both because they are generally unaware of what constitutes normal alertness and because sleepiness is generally viewed pejoratively, ascribed more often to a deficit in motivation than to an inadequately addressed physiologic sleep need.

Specific questioning about the occurrence of sleep episodes during normal waking hours, both intentional and unintentional, is necessary to determine the extent of the adverse effects of sleepiness on a patient’s daytime function. Specific areas to be addressed include the occurrence of inadvertent sleep episodes while driving or in other safety-related settings, sleepiness while at work or school (and the relationship of sleepiness to work and school performance), and the effect of sleepiness on social and family life. Standardized questionnaires, e.g., the Epworth Sleepiness Scale, are now commonly used in clinical and research settings to quantify daytime sleep tendency and screen for excessive sleepiness.

Driving is particularly hazardous for patients with increased sleepiness. Reaction time is equally impaired by 24 h of sleep loss as by a blood alcohol level of 0.10 g/dL. More than half of Americans admit to having fallen asleep while driving. An estimated 250,000 motor vehicle crashes per year are due to drowsy drivers, causing about 20% of all serious crash injuries and deaths. Drowsy driving legislation, aimed at improving education of all drivers about the hazards of driving drowsy and establishing sanctions comparable to those for drunk driving, has been enacted in New Jersey and is pending in several other states. Screening for sleep disorders, provision of an adequate number of safe highway rest areas, maintenance of unobstructed shoulder rumble strips, and strict enforcement and compliance monitoring of hours-of-service policies are needed to reduce the risk of sleep-related transportation crashes. Evidence for significant daytime impairment in association either with the diagnosis of a primary sleep disorder, such as narcolepsy or sleep apnea, or with imposed or self-selected sleep-wake schedules (see “Shift-Work Disorder”) raises the issue of the physician’s responsibility to notify motor vehicle licensing authorities of the increased risk of sleepiness-related motor vehicle crashes. As with epilepsy, legal requirements vary from state to state, and existing legal precedents do not provide a consistent interpretation of the balance between the physician’s responsibility and the patient’s right to privacy. At a minimum, physicians should inform patients who report a history of nodding off or falling asleep at the wheel or who have excessive daytime sleepiness about the increased risk of operating a motor vehicle, advise such patients not to drive a motor vehicle until the cause of the excessive sleepiness has been diagnosed and successful treatment has been implemented, and reevaluate the patient to determine when it is safe for the patient to resume driving. Each of those steps should be documented in the patient’s medical record.

The distinction between fatigue and sleepiness can be useful in the differentiation of patients with complaints of fatigue or tiredness in the setting of disorders such as fibromyalgia, chronic fatigue syndrome (Chap. 52), or endocrine deficiencies such as hypothyroidism or Addison’s disease. While patients with these disorders can typically distinguish their daytime symptoms from the sleepiness that occurs with sleep deprivation, substantial overlap can occur. This is particularly true when the primary disorder also results in chronic sleep disruption (e.g., sleep apnea in hypothyroidism) or in abnormal sleep (e.g., fibromyalgia).

While clinical evaluation of the complaint of excessive sleepiness is usually adequate, objective quantification is sometimes necessary. Assessment of daytime functioning as an index of the adequacy of sleep can be made with the multiple sleep latency test (MSLT), which utilizes repeated measurement of sleep latency (time to onset of sleep) under standardized conditions during a day following quantified nocturnal sleep. The average latency across four to six tests (administered every 2 h across the waking day) provides an objective measure of daytime sleep tendency. Disorders of sleep that result in pathologic daytime somnolence can be reliably distinguished with the MSLT. In addition, the multiple measurements of sleep onset may identify direct transitions from wakefulness to REM sleep that are suggestive of specific pathologic conditions (e.g., narcolepsy).


Narcolepsy is both a disorder of the ability to sustain wakefulness voluntarily and a disorder of REM sleep regulation (Table 20-2). The classic “narcolepsy tetrad” consists of excessive daytime somnolence plus three specific symptoms related to an intrusion of REM sleep characteristics (e.g., muscle atonia, vivid dream imagery) into the transition between wakefulness and sleep: (1) sudden weakness or loss of muscle tone without loss of consciousness, often elicited by emotion (cataplexy); (2) hallucinations at sleep onset (hypnagogic hallucinations) or upon awakening (hypnopompic hallucinations); and (3) muscle paralysis upon awakening (sleep paralysis). The severity of cataplexy varies, as patients may have two to three attacks per day or per decade. Some patients with objectively confirmed narcolepsy (see later) may show no evidence of cataplexy. In those with cataplexy, the extent and duration of an attack may also vary, from a transient sagging of the jaw lasting a few seconds to rare cases of flaccid paralysis of the entire voluntary musculature for up to 20–30 min. Symptoms of narcolepsy typically begin in the second decade, although the onset ranges from ages 5–50 years. Once established, the disease is chronic without remissions. Secondary forms of narcolepsy have been described (e.g., after head trauma).

TABLE 20-2



Narcolepsy affects about 1 in 4000 people in the United States and appears to have a genetic basis. Recently, several convergent lines of evidence suggest that the hypothalamic neuropeptide hypocretin (orexin) is involved in the pathogenesis of narcolepsy: (1) a mutation in the hypocretin receptor 2 gene has been associated with canine narcolepsy; (2) hypocretin “knockout” mice that are genetically unable to produce this neuropeptide exhibit behavioral and electrophysiologic features resembling human narcolepsy; and (3) cerebrospinal fluid levels of hypocretin are reduced in most patients who have narcolepsy with cataplexy. The inheritance pattern of narcolepsy in humans is more complex than in the canine model. However, almost all narcoleptics with cataplexy are positive for HLA DQB1*0602, suggesting that an autoimmune process may be responsible.


The diagnostic criteria continue to be a matter of debate. Certainly, objective verification of excessive daytime somnolence, typically with MSLT mean sleep latencies <8 min, is an essential if nonspecific diagnostic feature. Other conditions that cause excessive sleepiness, such as sleep apnea or chronic sleep deprivation, must be rigorously excluded. The other objective diagnostic feature of narcolepsy is the presence of REM sleep in at least two of the naps during the MSLT. Abnormal regulation of REM sleep is also manifested by the appearance of REM sleep immediately or within minutes after sleep onset in 50% of narcoleptic patients, a rarity in unaffected individuals maintaining a conventional sleep-wake schedule. The REM-related symptoms of the classic narcolepsy tetrad are variably present. There is increasing evidence that narcoleptics with cataplexy (one-half to two-thirds of patients) may represent a more homogeneous group than those without this symptom. However, a history of cataplexy can be difficult to establish reliably. Hypnagogic and hypnopompic hallucinations and sleep paralysis are often found in nonnarcoleptic individuals and may be present in only one-half of narcoleptics. Nocturnal sleep disruption is commonly observed in narcolepsy but is also a nonspecific symptom. Similarly, a history of “automatic behavior” during wakefulness (a trancelike state during which simple motor behaviors persist) is not specific for narcolepsy and serves principally to corroborate the presence of daytime somnolence.

TREATMENT Narcolepsy

The treatment of narcolepsy is symptomatic. Somnolence is treated with wake-promoting therapeutics. Modafinil is now the drug of choice, principally because it is associated with fewer side effects than older stimulants and has a long half-life; 200–400 mg is given as a single daily dose. Older drugs such as methylphenidate (10 mg bid to 20 mg qid) or dextroamphetamine (10 mg bid) are still used as alternatives, particularly in refractory patients.

These latter medications are now available in slow-release formulations, extending their duration of action and allowing once-daily dosing.

Treatment of the REM-related phenomena of cataplexy, hypnagogic hallucinations, and sleep paralysis requires the potent REM sleep suppression produced by antidepressant medications. The tricyclic antidepressants (e.g., protriptyline [10–40 mg/d] and clomipramine [25–50 mg/d]) and the selective serotonin reuptake inhibitors (SSRIs) (e.g., fluoxetine [10–20 mg/d]) are commonly used for this purpose. Efficacy of the anti-depressants is limited largely by anticholinergic side effects (tricyclics) and by sleep disturbance and sexual dysfunction (SSRIs). Alternately, gamma hydroxybutyrate (GHB), given at bedtime, and 4 h later, is effective in reducing daytime cataplectic episodes. Adequate nocturnal sleep time and planned daytime naps (when possible) are important preventive measures.


Respiratory dysfunction during sleep is a common, serious cause of excessive daytime somnolence as well as of disturbed nocturnal sleep. An estimated 2–5 million individuals in the United States have a reduction or cessation of breathing for 10–150 s from 30 to several hundred times every night during sleep. These episodes may be due to either an occlusion of the airway (obstructive sleep apnea), absence of respiratory effort (central sleep apnea), or a combination of these factors (mixed sleep apnea) (Fig. 20-3). Failure to recognize and treat these conditions appropriately may lead to impairment of daytime alertness, increased risk of sleep-related motor vehicle accidents, hypertension and other serious cardiovascular complications, and increased mortality. Sleep apnea is particularly prevalent in overweight men and in the elderly, yet it is estimated to remain undiagnosed in 80–90% of affected individuals. This is unfortunate since effective treatments are available.


The term parasomnia refers to abnormal behaviors or experiences that arise from or occur during sleep. A continuum of parasomnias arise from NREM sleep, from brief confusional arousals to sleepwalking and night terrors. The presenting complaint is usually related to the behavior itself, but the parasomnias can disturb sleep continuity or lead to mild impairments in daytime alertness. Two main parasomnias occur in REM sleep: REM sleep behavior disorder (RBD), which will be described later in this chapter, and nightmare disorder.

Sleepwalking (somnambulism)

Patients affected by this disorder carry out automatic motor activities that range from simple to complex. Individuals may walk, urinate inappropriately, eat, or exit from the house while remaining only partially aware. Full arousal may be difficult, and individuals may rarely respond to attempted awakening with agitation or even violence. Sleepwalking arises from slow-wave sleep (NREM stage N3 sleep), usually in the first 2 h of the night, and is most common in children and adolescents, when these sleep stages are most robust. Episodes are usually isolated but may be recurrent in 1–6% of patients. The cause is unknown, though it has a familial basis in roughly one-third of cases.

Sleep terrors

This disorder, also called pavor nocturnus, occurs primarily in young children during the first several hours after sleep onset, in slow-wave sleep (NREM stage N3 sleep). The child suddenly screams, exhibiting autonomic arousal with sweating, tachycardia, and hyperventilation. The individual may be difficult to arouse and rarely recalls the episode on awakening in the morning. Parents are usually reassured to learn that the condition is self-limited and benign and that no specific therapy is indicated. Both sleep terrors and sleepwalking represent abnormalities of arousal. In contrast, nightmares occur during REM sleep and cause full arousal, with intact memory for the unpleasant episode.

Sleep bruxism

Bruxism is an involuntary, forceful grinding of teeth during sleep that affects 10–20% of the population. The patient is usually unaware of the problem. The typical age of onset is 17–20 years, and spontaneous remission usually occurs by age 40. Sex distribution appears to be equal. In many cases, the diagnosis is made during dental examination, damage is minor, and no treatment is indicated. In more severe cases, treatment with a rubber tooth guard is necessary to prevent disfiguring tooth injury. Stress management or, in some cases, biofeedback can be useful when bruxism is a manifestation of psychological stress. There are anecdotal reports of benefit using benzodiazepines.

Sleep enuresis

Bedwetting, like sleepwalking and night terrors, is another parasomnia that occurs during sleep in the young. Before age 5 or 6 years, nocturnal enuresis should probably be considered a normal feature of development. The condition usually improves spontaneously by puberty, has a prevalence in late adolescence of 1–3%, and is rare in adulthood. In older patients with enuresis, a distinction must be made between primary and secondary enuresis, the latter being defined as bed-wetting in patients who have previously been fully continent for 6–12 months. Treatment of primary enuresis is reserved for patients of appropriate age (>5 or 6 years) and consists of bladder training exercises and behavioral therapy. Urologic abnormalities are more common in primary enuresis and must be assessed by urologic examination. Important causes of secondary enuresis include emotional disturbances, urinary tract infections or malformations, cauda equina lesions, epilepsy, sleep apnea, and certain medications. Symptomatic pharmacotherapy is usually accomplished with desmopressin (0.2 mg qhs), oxybutynin chloride (5–10 mg qhs), or imipramine (10–50 mg qhs).

Miscellaneous parasomnias

Other clinical entities may be characterized as a parasomnia or a sleep-related movement disorder in that they occur selectively during sleep and are associated with some degree of sleep disruption. Examples include jactatio capitis nocturna (nocturnal headbanging, rhythmic movement disorder), confusional arousals, sleep-related eating disorder, and nocturnal leg cramps.

REM sleep behavior disorder (RBD)

RBD is a rare condition that is distinct from other parasomnias in that it occurs during REM sleep. It primarily afflicts men of middle age or older, many of whom have an existing, or developing, neurologic disease. Approximately one-half of patients with RBD will develop Parkinson’s disease (Chap. 30) within 10–20 years. Presenting symptoms consist of agitated or violent behavior during sleep, as reported by a bed partner. In contrast to typical somnambulism, injury to the patient or bed partner is not uncommon, and, upon awakening, the patient reports vivid, often unpleasant, dream imagery. The principal differential diagnosis is nocturnal seizures, which can be excluded with polysomnography. In RBD, seizure activity is absent on the EEG, and disinhibition of the usual motor atonia is observed in the EMG during REM sleep, at times associated with complex motor behaviors. The pathogenesis is unclear, but damage to brainstem areas mediating descending motor inhibition during REM sleep may be responsible. In support of this hypothesis are the remarkable similarities between RBD and the sleep of animals with bilateral lesions of the pontine tegmentum in areas controlling REM sleep motor inhibition. Treatment with clonazepam (0.5–1.0 mg qhs) provides sustained improvement in almost all reported cases.


A subset of patients presenting with either insomnia or hypersomnia may have a disorder of sleep timing rather than sleep generation. Disorders of sleep timing can be either organic (i.e., due to an abnormality of circadian pacemaker[s] or its input from entraining stimuli) or environmental (i.e., due to a disruption of exposure to entraining stimuli from the environment). Regardless of etiology, the symptoms reflect the influence of the underlying circadian pacemaker on sleep-wake function. Thus, effective therapeutic approaches should aim to entrain the oscillator at an appropriate phase.

Jet lag disorder

More than 60 million persons experience transmeridian air travel annually, which is often associated with excessive daytime sleepiness, sleep onset insomnia, and frequent arousals from sleep, particularly in the latter half of the night. Gastrointestinal discomfort is common. The syndrome is transient, typically lasting 2–14 d depending on the number of time zones crossed, the direction of travel, and the traveler’s age and phase-shifting capacity. Travelers who spend more time outdoors reportedly adapt more quickly than those who remain in hotel rooms, presumably due to brighter (outdoor) light exposure. Avoidance of antecedent sleep loss and obtaining nap sleep on the afternoon prior to overnight travel greatly reduce the difficulty of extended wakefulness. Laboratory studies suggest that sub milligram doses of the pineal hormone melatonin can enhance sleep efficiency, but only if taken when endogenous melatonin concentrations are low (i.e., during biologic daytime), and that melatonin may induce phase shifts in human rhythms. A large-scale clinical trial evaluating the safety and efficacy of melatonin as a treatment for jet lag disorder and other circadian sleep disorders is needed.

In addition to jet lag associated with travel across time zones, many patients report a behavioral pattern that has been termed social jet lag, in which their bedtimes and wake times on weekends or days off occur 4–8 h later than they do during the week. This recurrent displacement of the timing of the sleep-wake cycle is common in adolescents and young adults and is associated with sleep onset insomnia, poorer academic performance, increased risk of depressive symptoms, and excessive daytime sleepiness.

Shift-work disorder

More than 7 million workers in the United States regularly work at night, either on a permanent or rotating schedule. Many more begin work between 4 A.M. and 7 A.M., requiring them to commute and then work during the time of day that they would otherwise be asleep. In addition, each week millions more elect to remain awake at night to meet deadlines, drive long distances, or participate in recreational activities. This results in both sleep loss and misalignment of the circadian rhythm with respect to the sleep-wake cycle.

Studies of regular night-shift workers indicate that the circadian timing system usually fails to adapt successfully to such inverted schedules. This leads to a misalignment between the desired work-rest schedule and the output of the pacemaker and in disturbed daytime sleep in most individuals. Sleep deprivation, increased length of time awake prior to work, and misalignment of circadian phase produce decreased alertness and performance, increased reaction time, and increased risk of performance lapses, thereby resulting in greater safety hazards among night workers and other sleep-deprived individuals. Sleep disturbance nearly doubles the risk of a fatal work accident. Additional problems include higher rates of breast, colorectal, and prostate cancer and of cardiac, gastrointestinal, and reproductive disorders in long-term night-shift workers. Recently, the World Health Organization has added night-shift work to its list of probable carcinogens.

Sleep onset is associated with marked attenuation in perception of both auditory and visual stimuli and lapses of consciousness. The sleepy individual may thus attempt to perform routine and familiar motor tasks during the transition state between wakefulness and sleep (stage N1 sleep) in the absence of adequate processing of sensory input from the environment. Motor vehicle operators are especially vulnerable to sleep-related accidents since the sleep-deprived driver or operator often fails to heed the warning signs of fatigue. Such attempts to override the powerful biologic drive for sleep by the sheer force of will can yield a catastrophic outcome when sleep processes intrude involuntarily upon the waking brain. Such sleep-related attentional failures typically last only seconds but are known on occasion to persist for longer durations. These frequent brief intrusions of stage N1 sleep into behavioral wakefulness are a major component of the impaired psychomotor performance seen with sleepiness. There is a significant increase in the risk of sleep-related, fatal-to-the-driver highway crashes in the early morning and late afternoon hours, coincident with bimodal peaks in the daily rhythm of sleep tendency.

Resident physicians constitute another group of workers at risk for accidents and other adverse consequences of lack of sleep and misalignment of the circadian rhythm. Recurrent scheduling of resident physicians to work shifts of 24 h or more consecutive hours impairs psychomotor performance to a degree that is comparable to alcohol intoxication, doubles the risk of attentional failures among intensive care unit interns working at night, and significantly increases the risk of serious medical errors in intensive care units, including a fivefold increase in the risk of serious diagnostic mistakes. Some 20% of hospital interns report making a fatigue-related mistake that injured a patient, and 5% admit making a fatigue-related mistake that results in the death of a patient. Moreover, working for >24 h consecutively increases the risk of percutaneous injuries and more than doubles the risk of motor vehicle crashes on the commute home. For these reasons, in 2008 the Institute of Medicine concluded that the practice of scheduling resident physicians to work for more than 16 consecutive hours without sleep is hazardous for both resident physicians and their patients.

From 5 to 10% of individuals scheduled to work at night or in the early morning hours have much greater than average difficulties remaining awake during night work and sleeping during the day; these individuals are diagnosed with chronic and severe shift-work disorder (SWD). Patients with this disorder have a level of excessive sleepiness during night work and insomnia during day sleep that the physician judges to be clinically significant; the condition is associated with an increased risk of sleep-related accidents and with some of the illnesses associated with night-shift work. Patients with chronic and severe SWD are profoundly sleepy at night. In fact, their sleep latencies during night work average just 2 min, comparable to mean sleep latency durations of patients with narcolepsy or severe daytime sleep apnea.

TREATMENT Shift-Work Disorder

Caffeine is frequently used to promote wakefulness. However, it cannot forestall sleep indefinitely, and it does not shield users from sleep-related performance lapses. Postural changes, exercise, and strategic placement of nap opportunities can sometimes temporarily reduce the risk of fatigue-related performance lapses. Properly timed exposure to bright light can facilitate rapid adaptation to night-shift work.

While many techniques (e.g., light treatment) used to facilitate adaptation to night-shift work may help patients with this disorder, modafinil is the only therapeutic intervention that has ever been evaluated as a treatment for this specific patient population. Modafinil (200 mg, taken 30–60 min before the start of each night shift) is approved by the U.S. Food and Drug Administration as a treatment for the excessive sleepiness during night work in patients with SWD. Although treatment with modafinil significantly increases sleep latency and reduces the risk of lapses of attention during night work, SWD patients remain excessively sleepy at night, even while being treated with modafinil.

Safety programs should promote education about sleep and increase awareness of the hazards associated with night work. The goal should be to minimize both sleep deprivation and circadian disruption. Work schedules should be designed to minimize (1) exposure to night work, (2) the frequency of shift rotation so that shifts do not rotate more than once every 2–3 weeks, (3) the number of consecutive night shifts, and (4) the duration of night shifts. Shift durations of >16 h should be universally recognized as increasing the risk of sleep-related errors and performance lapses to a level that is unacceptable in nonemergency circumstances. At least 11 h off duty should be provided between work shifts, with at least 1 day off every week and 2 consecutive days off every month. Additional off duty time should be allocated after night work, since sleep efficiency is much lower during daytime hours.

Delayed sleep phase disorder

Delayed sleep phase disorder is characterized by (1) reported sleep onset and wake times intractably later than desired, (2) actual sleep times at nearly the same clock hours daily, and (3) essentially normal all-night polysomnography except for delayed sleep onset. Patients exhibit an abnormally delayed endogenous circadian phase, with the temperature minimum during the constant routine occurring later than normal. This delayed phase could be due to (1) an abnormally long, genetically determined intrinsic period of the endogenous circadian pacemaker; (2) an abnormally reduced phase-advancing capacity of the pacemaker; (3) a slower rate of buildup of homeostatic sleep drive during wakefulness; or (4) an irregular prior sleep-wake schedule, characterized by frequent nights when the patient chooses to remain awake well past midnight (for social, school, or work reasons). In most cases, it is difficult to distinguish among these factors, since patients with an abnormally long intrinsic period are more likely to “choose” such late-night activities because they are unable to sleep at that time. Patients tend to be young adults. This self-perpetuating condition can persist for years and does not usually respond to attempts to reestablish normal bedtime hours. Treatment methods involving bright-light phototherapy during the morning hours or melatonin administration in the evening hours show promise in these patients, although the relapse rate is high.

Advanced sleep phase disorder

Advanced sleep phase disorder (ASPD) is the converse of the delayed sleep phase syndrome. Most commonly, this syndrome occurs in older people, 15% of whom report that they cannot sleep past 5 A.M., with twice that number complaining that they wake up too early at least several times per week. Patients with ASPD experience excessive daytime sleepiness during the evening hours, when they have great difficulty remaining awake, even in social settings. Typically, patients awaken from 3 to 5 A.M. each day, often several hours before their desired wake times. In addition to age-related ASPD, an early-onset familial variant of this condition has also been reported. In one such family, autosomal dominant ASPD was due to a missense mutation in a circadian clock component (PER2, as shown in Fig. 20-2) that altered the circadian period. Patients with ASPD may benefit from bright-light phototherapy during the evening hours, designed to reset the circadian pacemaker to a later hour.

Non-24-h sleep-wake disorder

This condition can occur when the synchronizing input (i.e., the light-dark cycle) from the environment to the circadian pacemaker is compromised (as in many blind people with no light perception) or when the maximal phase-advancing capacity of the circadian pacemaker is not adequate to accommodate the difference between the 24-h geophysical day and the intrinsic period of the pacemaker in the patient. Alternatively, patients’ self-selected exposure to artificial light may drive the circadian pacemaker to a >24-h schedule. Affected patients are not able to maintain a stable phase relationship between the output of the pacemaker and the 24-h day. Such patients typically present with an incremental pattern of successive delays in sleep propensity, progressing in and out of phase with local time. When the patient’s endogenous circadian rhythms are out of phase with the local environment, insomnia coexists with excessive daytime sleepiness. Conversely, when the endogenous circadian rhythms are in phase with the local environment, symptoms remit. The intervals between symptomatic periods may last several weeks to several months. Blind individuals unable to perceive light are particularly susceptible to this disorder, although it can occur in sighted patients. Nightly low-dose (0.5 mg) melatonin administration has been reported to improve sleep and, in some cases, to induce synchronization of the circadian pacemaker.


Prominent circadian variations have been reported in the incidence of acute myocardial infarction, sudden cardiac death, and stroke, the leading causes of death in the United States. Platelet aggregability is increased in the early morning hours, coincident with the peak incidence of these cardiovascular events. Misalignment of circadian phase, such as occurs during night-shift work, induces insulin resistance and higher glucose levels in response to a standard meal. Blood pressure of night workers with sleep apnea is higher than that of day workers. A better understanding of the possible role of circadian rhythmicity in the acute destabilization of a chronic condition such as atherosclerotic disease could improve the understanding of its pathophysiology.

Diagnostic and therapeutic procedures may also be affected by the time of day at which data are collected. Examples include blood pressure, body temperature, the dexamethasone suppression test, and plasma cortisol levels. The timing of chemotherapy administration has been reported to have an effect on the outcome of treatment. In addition, both the toxicity and effectiveness of drugs can vary during the day. For example, more than a fivefold difference has been observed in mortality rates following administration of toxic agents to experimental animals at different times of day. Anesthetic agents are particularly sensitive to time-of-day effects. Finally, the physician must be increasingly aware of the public health risks associated with the ever-increasing demands made by the duty-rest-recreation schedules in our round-the-clock society.